Methods and apparatuses for user equipment selecting and scheduling in intelligent wireless system

Abstract

Disclosed are methods and apparatuses for user equipment (UE) selecting and scheduling in an intelligent wireless system. An embodiment of the subject application provides a base station (BS). The BS includes a processor and a wireless transceiver coupled to the processor. The processor is configured to: obtain a number N and a first channel gain threshold, wherein the number N and the first channel gain threshold are determined based at least in part on uplink channel state information between the BS and multiple UEs; transmit, with the wireless transceiver, a scheduling indicator report configuration to each of the multiple UEs; receive, with the wireless transceiver, multiple scheduling indicators; and select the number N of UEs for participating in local model training according to the multiple scheduling indicators.

1 . A base station (BS), comprising:
 at least one memory; and   at least one processor coupled with the at least one memory and configured to cause the BS to:
 obtain a number N and a first channel gain threshold, wherein the number N and the first channel gain threshold are determined based at least in part on uplink channel state information between the BS and multiple user equipments (UEs); 
 transmit, to each of the multiple UEs, a scheduling indicator report configuration; 
 receive multiple scheduling indicators; and 
 select the number N of the multiple UEs to participate in local model training according to the multiple scheduling indicators. 
   
     
     
         2 . The BS of  claim 1 , wherein the at least one processor is configured to cause the BS to obtain the number N and the first channel gain threshold in response to at least one of:
 a new global model being applied;   a global model convergence being achieved; or   a convergence speed of a global model being lower than a desired speed.   
     
     
         3 . The BS of  claim 1 , wherein to obtain the number N and the first channel gain threshold, the at least one processor is configured to cause the BS to determine the number N based on the uplink channel state information and a historical number of iterations for convergence of a global model, and determine the first channel gain threshold based on the uplink channel state information. 
     
     
         4 . The BS of  claim 1 , wherein to obtain the number N and the first channel gain threshold, the at least one processor is configured to cause the BS to:
 transmit, to a server, the uplink channel state information; and   receive, from the server, the first channel gain threshold and the number N.   
     
     
         5 . The BS of  claim 1 , wherein the scheduling indicator report configuration transmitted to a UE comprises at least one of:
 a resource for reporting a scheduling indicator calculated by the UE;   the first channel gain threshold; or   a report quantity indicating a report of the scheduling indicator of the UE.   
     
     
         6 . The BS of  claim 1 , wherein the scheduling indicator report configuration is transmitted via one of:
 radio resource control (RRC) signaling;   a medium access control (MAC) control element (CE);   downlink control information (DCI); or   artificial intelligence related signaling.   
     
     
         7 . The BS of  claim 1 , wherein to select the number N of the multiple UEs to participate in the local model training, the at least one processor is configured to cause the BS to select the number N of the multiple UEs with smallest scheduling indicator values among the multiple scheduling indicators. 
     
     
         8 . The BS of  claim 1 , wherein the at least one processor is further configured to cause the BS to:
 transmit, to the number N of the multiple UEs, a local model report configuration; and   receive, from the number N of the multiple UEs, updated local models.   
     
     
         9 . The BS of  claim 8 , wherein the local model report configuration transmitted to a UE comprises at least one of:
 a resource for reporting an updated local model of the UE;   a global model; or   a report quantity indicating a report of the updated local model of the UE.   
     
     
         10 . The BS of  claim 8 , wherein the at least one processor is further configured to cause the BS to:
 update a global model according to the updated local models; and   if convergence of the updated global model is not achieved, then transmit, to the number N of the multiple UEs, the local model report configuration containing the updated global model for further local model training.   
     
     
         11 . The BS of  claim 8 , wherein the at least one processor is further configured to cause the BS to:
 receive, from a server, a local model report configuration trigger message comprising at least one of:
 a maximum latency for reporting local models; 
 identifiers (IDs) of the multiple UEs; 
 a report quantity indicating a report of an updated local model; 
 the number N; 
 an interim report quantity indicating a report of a schedule indicator; 
 a global model; or 
 the first channel gain threshold; and 
   transmit, to the server, the updated local models which satisfy the maximum latency.   
     
     
         12 . The BS of  claim 8 , wherein the at least one processor is further configured to cause the BS to:
 update a global model according to the updated local models; and   if convergence of the updated global model is not achieved:
 determine a second threshold for a scheduling indicator according to previously received scheduling indicators; 
 transmit, to the number N of the multiple UEs, a quantized scheduling indicator report configuration; 
 receive multiple quantized scheduling indicators; and 
 re-select the number N of the multiple UEs to participate in the local model training according to the multiple quantized scheduling indicators. 
   
     
     
         13 . The BS of  claim 11 , wherein the at least one processor is further configured to cause the BS to:
 receive, from the server, a second channel gain threshold for the scheduling indicator;   transmit, to each of the multiple UEs, a quantized scheduling indicator report configuration;   receive multiple quantized scheduling indicators; and   re-select the number N of the multiple UEs to participate in the local model training according to the multiple quantized scheduling indicators.   
     
     
         14 . A user equipment (UE) comprising:
 at least one memory; and   at least one processor coupled with the at least one memory and configured to cause the UE to:
 receive a scheduling indicator report configuration comprising at least one of a first resource for reporting a scheduling indicator calculated by the UE, a first channel gain threshold, or a report quantity indicating a report of the scheduling indicator of the UE; 
 calculate the scheduling indicator based at least in part on the first channel gain threshold; and 
 transmit the calculated scheduling indicator on the first resource. 
   
     
     
         15 . A server comprising:
 at least one memory; and   at least one processor coupled with the at least one memory and configured to cause the server to:
 receive uplink channel state information between multiple user equipments (UEs) and a base station (BS); 
 determine a first channel gain threshold and a number N based at least in part on the uplink channel state information, wherein the number N is a number of UEs to participate in local model training; and 
 transmit a local model report configuration trigger message comprising at least one of a report quantity indicating a report of an updated local model, the number N, an interim report quantity indicating a report of a scheduling indicator, the first channel gain threshold, a maximum latency for reporting local models, or a global model. 
   
     
     
         16 . The UE of  claim 14 , wherein the scheduling indicator report configuration comprising the first channel gain threshold is based at least in part on one or more of:
 a new global model being applied;   a global model convergence being achieved; or   a convergence speed of a global model being lower than a desired speed.   
     
     
         17 . The UE of  claim 14 , wherein the scheduling indicator report configuration is received via one of:
 radio resource control (RRC) signaling;   a medium access control (MAC) control element (CE);   downlink control information (DCI); or   artificial intelligence related signaling.   
     
     
         18 . The UE of  claim 14 , wherein the at least one processor is configured to cause the UE to receive a local model report configuration. 
     
     
         19 . The UE of  claim 18 , wherein the local model report configuration comprises at least one of:
 a resource for reporting an updated local model of the UE;   a global model; or   a report quantity indicating a report of the updated local model of the UE.   
     
     
         20 . A processor for wireless communication, comprising:
 at least one controller coupled with at least one memory and configured to cause the processor to:
 receive a scheduling indicator report configuration comprising at least one of a first resource for reporting a scheduling indicator calculated by a user equipment (UE), a first channel gain threshold, or a report quantity indicating a report of the scheduling indicator of the UE; 
 calculate the scheduling indicator based at least in part on the first channel gain threshold; and 
 transmit the calculated scheduling indicator on the first resource.